Building consumer confidence in AI at banks
85% of consumers say confidence in AI would grow with at least one guardrail in place
Trust, transparency and guardrails are key to building confidence around AI application at financial institutions.
That’s what TD Bank’s U.S. AI Insights Report, released March 31, found.
“Our survey reveals that 85% of consumers say their confidence in how banks use AI would be increased if we implemented just one guardrail,” Ted Paris, head of analytics, intelligence and AI at TD Bank U.S., told FinAi News, citing the report, which surveyed more than 2,500 customers. The survey was conducted by custom research solutions provider Big Village Feb. 18-25.
The following AI guardrails would provide confidence to bank customers, according to the survey:
- Human-in-the-loop oversight, 48%;
- Strong data security and customer privacy protection, 46%; and
- Transparency about AI application, 46%.
“At TD, we’ve adopted each of the mentioned guardrails and more to ensure that our clients can feel confident in how we’re using AI,” Paris said.
AI at TD
TD has deployed AI throughout the institution. Uuses cases include:
- A gen AI-driven virtual assistant within customer service centers; and
- TD AI Prism, a platform that predicts consumer needs and provides solutions.
“At TD we continually monitor our AI solutions for validation, interpretability and fairness,” Paris said, noting that the bank has human-in-the-loop in place, guardrails that limit input and output, and continuous fine-tuning and testing of models.
“Trust is a foundational aspect of our AI strategy,” Paris said.
Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas.
